README.md

metadata

license: apache-2.0

An UncensoredLLM with Reasoning, what more could you want?

• Note this is an experimental preview and we will continue improving based on feedback, so please good and bad feedback welcome. Leave it in the Community Comments and I will address your feedback asap.

• Refusal Alignment is pretty clear where we draw the line:

You might be asking yourself: Why Uncensored AI? Prompt Template the model is trained on:

"instruction,input,output": "The following user query requires an answer that is derived from the context given. It is imperative that the answer is not based on external knowledge or assumptions. Analyze the context thoroughly to extract the necessary information that will form the basis of a clear and accurate answer.<｜User｜>\n%instruction%\n%input%\n<｜Assistant｜>\n%output%",

We just launched our beta, LOOK INTO IT at https://uncensored.ai